It strikes me that the models are struggling with 'what happens next'.
I notice that GFS turns thing mild in the longer term, but I'm not so
sure, a lot may unfold over the weekend when the GFS v UKMO contest
will have been won, one way or the other. The models seem to have a
'human' gut feeling that keeps pushing things milder, but of cause
thats not true and is certainly unscientific :-)

I'm still not convinced I'll see snow in Southend-on-Sea next week
yet. Never lone settling on the ground.

Interesting days ahead.

Keith (Southend)

********************************
'Weather Home & Abroad'
http://www.southendweather.net
********************************
COL Station for Southend-on-Sea
http://www.wunderground.com/weathers...p?ID=IESSEXSO1
********************************
Reply to: kreh'at'southendweather'dot'net
All mail scanned for virus's using Norton 2003

On Wed, 21 Jan 2004 18:27:06 -0000, "PJB"
wrote:

Quite a difference in the midday UK Global Model vs GFS 12Z, not really a
surprise as there was a siginificant difference between them at 00Z too.

The difference can be traced back to a possible low that forms in the
western Atlantic at around the T+48 time frame. The GFS has consistently
forecast a low latitude wave to move through the ridge, emerge on the cold
side of the long wave trough, engage the left exit and deepen as it heads
over Biscay, eventually developing into a large cold low over NW continental
europe by mid week. The low is not present at all on the UKMO run, either at
00Z or 12Z, and fax charts were modified "someway" towards the GFS solution.
The 12Z has remained undevlopmental in regards to this low. It is
unfortunate that this low has such major downstream ramifications, with the
GFS having a very cold pool of air over NW Europe by midweek, vs MetO Global
which has a much more dominant high ridging east. The Issued T+72 Fax chart
has been modified away from the raw 72 hr prog to develop this low latitude
wave low.

Other models are somewhat lack lustre, Nogaps and Canadian, are similar in
someways to the GFS, but are not so developmental. German however albeit
slightly different as you would expect, broadly supports the GFS evolution
for the low, giving extra support to modifications to the GM output and to
the outcome portrayed by the 12Z GFS

It is interesting to compare the timeplot series for 850mb for london 12Z
run vs 00Z run, they are almost identical in the period T+60 to T+144. This
lends much more weight to the GFS solution which has shown great concistency
over the last 6 runs.